Cadreon Test Drives Google’s Cross-Device Measurement System

cadreonFor several months, IPG Mediabrands’ ad tech unit Cadreon has been testing Google’s cross-device measurement system, baked into DoubleClick advertiser products as of June.

Krish Sailam, senior product manager of mobile for Cadreon, said the trading desk has run a few beta tests with advertisers, and noted conversion uplift has consistently reached the double digits.

Cadreon is sharing early results from its first series of tests using Google’s new cross-device system – a campaign for a big automaker that ran from August until early October.

Client objectives included: measure the role of mobile in driving sales and lowering CPAs, optimize programmatic spend more effectively and understand how each channel (smartphone, desktop, tablet) contributed to a conversion.

Previously, insights on some of these performance guidelines were less than clear.

“Most of our targeting was based on probabilistic solutions and our attribution sources, whether that was the ad server or multitouch attribution tools, weren’t cross-device aware,” said Ian Johnson, EVP of global product for Cadreon. “The attribution credits we got back weren’t necessarily reflecting the gain we were actually getting.”

Cadreon began with the premise that automotive buyers begin the research process early before they commit to a purchase.

Although cars are a big-ticket item with the final transaction often occurring in-store, Cadreon’s client wanted to measure the impact of mobile along the purchase path. A consumer may be served a piece of media on one device, but may take an action on another.

Cadreon’s respective auto client used DoubleClick Campaign Manager as its ad server, so it made sense to use its device graph to help draw some of those conclusions. Cadreon first took the siloed approach – it pulled data from individual channels like smartphones, desktop and tablets to build a baseline conversion measurement.

Then it layered in cross-device data for both programmatic and reserved buys, across search and display, which “enabled us to dig down and show that a person who was shown an ad on their smartphone ultimately finished a lower-funnel action on the desktop,” Sailam added.

Initially, Cadreon quantified the impact of mobile as a single-channel. But this didn’t take into account the total cost of conversion, regardless of channel impact.

So using DoubleClick’s cross-device conversion metrics, it ran a series of tests that would help it “fact check” mobile’s real impact. In one instance, Cadreon maintained a fixed budget, meaning it varied the percentage allocated to mobile. In the other instance, the total budget increased with all new spend going to mobile.

Tapping into conversion metrics enabled Cadreon to refine mobile audience segments and determine the impact of mobile or tablet impressions on desktop media. Ultimately, it got a clearer picture on attribution.

“In many cases, we saw data that suggested people were using multiple desktops to complete one action,” Sailam added, “so there was typically a work and home computer used by a single person.”

By optimizing buys based on cross-channel activity (e.g. serving more desktop ads in the after-work hours because a person falls into a segment that prefers to shop and surf online at night) Cadreon's client saw a 15% lift in overall conversions, a 22% increase in conversions when smartphones were a measured touch point in the purchase funnel, and a 13% reduction in eCPA.

Cross device measurement gives clients more of a holistic view into their media spend, claims Sailam.

Although Facebook’s people-based system is arguably more robust right now in terms of its persistent, logged-in user base (Google’s logged-in userset is much smaller, though it applies anonymized data from previously logged-in users to build probabilistic models based on conversion/device type and location), Cadreon’s Johnson said the Facebook Atlas bidder (editor: Atlas is Facebook's cross-device measurement system/ad server) competes with some of its main buying platforms.

And, Sailam has questions about its breadth of measurement cross-channel.

“One of the things that concerns me about Facebook is that their audience is going mobile so quickly that you start to lose some of the cross-device aspect,” he said. “The Facebook graph might indicate users on tablets and smartphones really well, but their presence on desktop isn’t measured so much. Going into next year, Google is talking about different services that will … ensure they’re the largest both in terms of scale but also keeping that logged-in data set very fresh.”

1 Comment

  1. While this sounds like a great breakthrough, I wonder if they were able to shed some light on how they actually measured a conversion with an automotive client (as noted, being the purchase of a vehicle in store) and how they linked that to a device.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>